AutoVAI
to compute formant-based articulation features from read speech
Install / Use
/learn @SPEECHCOG/AutoVAIREADME
LIU Yuanyuan (yuanyuan.liu@tuni.fi), TUT, 2021-3-29.
In thi folder, the least programs used for automatic computation of vowel articulation features are included.
To use this tool:
- environment of python3.
- the phoneme recognizer Allosaurus should be installed: "pip install ipa_recognizer".
- software praat should be installed.
- In run.py, the directory for praat should be customized.
- In the two praat scripts used in run.py, the value of filepath_source should be customized.
Citation of this work: Yuanyuan Liu, Nelly Penttilä, Tiina Ihalainen, Juulia Lintula, Rachel Convey, Okko Räsänen, “Language-Independent Approach for Automatic Computation of Vowel Articulation Features in Dysarthric Speech Assessment”, IEEE/ACM Transactions on Audio, Speech and Language Processing, 15 June, 2021. DOI: 10.1109/TASLP.2021.3090973.
